The long-term goals of this project are to better understand, and to improve upon, the structural characterization and refinement processes for macromolecules by modifying new structure-improvement techniques (developed and proven effective for crystal structures) for application to Nuclear Magnetic Resonance Spectroscopy (NMR) structures. More accurate, detailed structures from either method will help in understanding biological function and aid the drug design and protein design work in numerous medical applications. The approach that will be employed is to use real-time-updatable visualizations that emphasize any poor matches of NMR distance restraints, all-atom steric contacts, backbone phi, psi angles, sidechain rotamers, and other geometry versus the current local model to enable either correction of individual conformations or definition of new restraints for further ensemble refinement, thus maximizing agreement both to data and other criteria. Methodology will be tested during refinement of new NMR structures, extended to other NMR experimental data types, and automated as far as feasible. All results, including any new methodologies, will be published and provided as a web service for use by students and researchers.